We propose to use mouse models of breast, pancreatic islets, and ovarian tumors to investigate tumor progression, with emphasis on the genetic events that influence spread of primary tumors to lymph nodes and distant organs. (1) We will initially survey a variety of genetically manipulated mouse strains known to develop primary carcinomas of the breast, ovary, and pancreas to gauge the time, location, and pathological features of tumor spread beyond the primary lesions, without further experimental perturbation. (2) Several methods will then be used to ascertain whether known genes promote progression in different models and to attempt to discover genes that affect progression. These methods include: (a) introduction of specific coding sequences for known proteins or cDNA libraries of varying complexity into premalignant lesions or primary tumors by infection with avian retrovirus vectors; (b) array-based comparative genome hybridization (CGH) to look for genes amplified and deleted in the progression of tumors from primary to metastatic states; and (c) cDNA microarray methods to compare gene expression patterns in different stages of mammary oncogenesis in MMTV-Wnt-1 transgenics with or without loss of Pten or P53 genes. (3) In addition, we will try to determine whether passage of tumor cells through lymph nodes influences pulmonary or hepatic metastasis by using lymphotoxin-beta fusions to block formation of lymph nodes during gestation and by attempting to mark tumor cells that have been exposed to factors present in lymph nodes.